lecture10overviewandsummarization-教學(xué)講解課件

ResearchinMarketingAutumn2019OverviewoftopicsdiscussedDr.AtanuK.NathDepartmentofMarketingProductionvs.MarketorientationAproductionorientationIsinwardlookingFocusesonitsownproductionline-upDrumsupamarketthroughaggressivesellingResistanttochange–itthreatenstheline-uphenceresistanttomarketresearchExamples?SchoolofManagementsom.surrey.ac.ukWheredoesresearchinmarketingfitin?Formarketingorientation:DefinethegenericneedsDefinetargetgroupsandidentifywhichonestoserveSettleonadifferentialadvantagestrategyUsemarketresearchthenstartagainSchoolofManagementsom.surrey.ac.ukMarketingresearchis

achangeagentDefiningMR:“Marketingresearchistheprocessofdesigning,gathering,analyzing,andreportinginformationthatmaybeusedtosolveaspecificmarketingproblem.”–Burns,Bush,2019SchoolofManagementsom.surrey.ac.ukEssentialsofResearchinMarketingProblemidentificationEstablishresearchquestionIdentificationofappropriatemethodJustificationofmethodandapproachAcknowledgmentofthelimitationsofchosen,andalternativemethodsProviderelevantexamplesFigure1.5 Simplelineardescriptionofthemarketingresearchprocess

Whenshouldmarketingresearchbeconducted?Howwelldoesthefirmknowthemarket?Howmuchwilltheresearchcost?Willtheresearchaddvalue?Howlongwilltheresearchtake?Whatisthequalityoftheresearch?7RIM09AssessingInformationNeedsIsitrelevantandnecessary?decision-basedutilityavoidinformationoverloadIsitfeasibletoacquire?timecostavailability8RIM09Burns&Bush(2019)Chisnall(2019)Malhotra&Birks(2019)Pre-proposaldevelopmentProposaldevelopmentPost-proposaldevelopmentReportingResearchbriefResearchproposalDatacollectionDataanalysis&evaluationPreparation&presentationofresearchreportProblemdefinitionResearchapproachdevelopedResearchdesigndevelopedFieldworkordatacollectionDatapreparation&analysisReportpreparation&presentationRIM099SpecificdecisionareaswithineachstageEstablishifthereisaneedDefinetheproblempreciselyEstablishobjectivesMethodsofaccessingdataIdentifyinformationsourcesDetermineresearchdesignDatacollectionsforms/questionnairesDecideonsamplecontentandsizeCollectdataDrawconclusions,andfinalreportpresentationAnalyzethedatacollectedPre-proposaldevelopmentphaseProposaldevelopmentstagePost-proposaldevelopmentstageReportingSource:Burns&Bush(2019),p.2410RIM09Stage1-DefiningthemarketingproblemInmarketingresearch,aproblemDOESNOTnecessarilyrefertoatrouble.But,itcanalsomeananopportunity(e.g.marketdevelopment,brandextension,newserviceofferingetc)UnderstandingthemarketingdecisionproblemDefiningthemarketingresearchproblem11RIM09ClassificationsofresearchproblemMarketingdecisionproblemTheproblemconfrontingthemarketingdecision-maker,whichaskswhatthedecisionmakerneedstodoe.g.shouldtheadvertisingcampaignbechanged?MarketingresearchproblemAproblemthatentailsdeterminingwhatinformationisneededandhowitcanbeobtainedinthemostfeasiblewaye.g.whatistheeffectivenessofthecurrentadvertisingcampaign?RIM0912Stage1–Definingthemarketingproblem(cont)DevelopingasetofresearchobjectivesWhatspecificinformationdoweneedtoanswertheresearchproblem?Specifically,whatdowewanttofindoutinordertomakeadecision?13RIM09Marketingproblemsandobjectives:

AsynopsisMarketingresearchproblemMarketingresearchobjectiveMarketingdecisionproblemBroadly,whatinformationisneededandhowcanthatinformationbeobtainedefficientlyandeffectively?Informationoriented.Whatspecificpiecesofinformationareneededtosolvethemarketingresearchproblemandmakeadecision?

Abroader-basedproblemthatrequiresmarketingresearchinorderformanagerstotakeproperactions.Actionoriented.14RIM09Stage2:Planning&decidingtheresearchdesignResearchdesignspecifieswhichresearchquestionsmustbeanswered?howandwhenwillthedatabegathered?whoshouldbeinterviewedorsurveyed?howwillthedatabeanalysed?3typesofresearchdesignexploratorydescriptivecausal15RIM09ExploratoryresearchdesignOftena1ststepinmarketingresearchandprovidesinsightsintomarketingresearchproblemMakefulluseofsecondary(published)dataTendstousequalitativeassessmentsratherthandetailedquantitativedataArelativelyspeedy&economicwayofacquiringoverviewofproblemanditsrelevantfactors,usefulindevelopinghypothesesaboutspecificmarkets;Particularlyvaluableasa‘researchfilter’beforefurthercommitmentsmadetomoreextensiveandexpensiveresearchactivitiesRIM0916DescriptiveresearchdesignGenerallydevelopedfromexploratoryresearchfindings;describesspecificmarketphenomena(e.g.productusage,frequencyofstorepatronageetc)CollectsstatisticaldatatotesthypothesesdevelopedfromexploratoryresearchProvidesdataforcomparativeanalysesofcompetitiveproductsDevelopsprofilesoftypesofcustomersandtheirpreferencesRIM0917CausalresearchdesignAttemptstoidentifycause-and-effectrelationshipsbetweenspecificmarketbehaviour(e.g.advertisingexpendituresleadtosales)Butcorrelationsamongvariablesneedtobeevaluatedcautiously–associationisnotnecessarilycausationTendstouseexperimentationsothatcertainfactorscanbecontrolledandcomparedRIM0918AnalysingthedataCoding(preandpostcoding)EditingDataentrycheckingforerrorsRunningfrequenciescheckingfornormality(skewnessetc.)Crosstabulationsrelationshipsanddifferencesamongfactorse.g.usagebygender,attitudebyage19RIM09Analysingthedata(cont)EstablishingthereliabilityandvalidityofthedataReliability=thestabilityandconsistencyoftheresultsderivedfromtheresearchRIM0920Analysingthedata(cont)Validity=howwellaspecificresearchmethodmeasureswhatitclaimstomeasureInternalvalidityarethemeasuresrelatedtoaresearchrigorouslydeveloped?constructvalidity,contentvalidity,concurrentvalidity&predictivevalidityFacevaliditydotheresultsappearplausibleorlogicalinthelackofsupportingevidence?Externalvaliditycantheresultsbegeneralisabletootherdissimilarresearchsituations?RIM0921ReportingtheresultsReportconclusionsandrecommendationsPresentingthedataaddressingeachresearchobjectivereporttheresearchmethodusedexplaintheanalysisconductedusetablesandchartsforeaseofinterpretationclarityinwritingconcise22RIM095mainstagesofmarketingresearchProblemdefinitionResearchdesignDatacollectionDatapreparationandanalysisReportpreparationandpresentationEachstageinvolvesmultipledecisionareasRIM0923SamplingSamplingAsampleisasubsetofpopulation.Thepatternobservedinasamplecanbereplicatedinthepopulation.Inotherwords,thefindingsfromasamplecanbegeneralizedtothepopulation.Agoodsamplehastwocharacteristics:AccuracywhichreferstoasampleinwhichthereislittleornobiasPrecisionwhichreferstoasamplethathasanacceptablelevelofsamplingerror.SamplingprocessChoosingthesamplingframe:Howshouldthepopulationbedetermined?Samplingframeisacompletelistofelementsinthepopulationfromwhichsampleisdrawn.Examplesofsamplingframesare:Theyellowpageslistingofrestaurants,thetelephonedirectorylistingofindividual,acompany’sinternaldatabaselistingitsemployeesorcustomers,electronicdirectoriesavailableoninternet,universityregistrationlists.SamplingprocessSelectingthesamplingmethod:Howshouldaresearchergoaboutdrawingthesample?ProbabilitysamplingThistypeofsamplingisbasedonrandomselectionandaparticularpopulationelementorunitwillbeincludedinthesampleisknown.Thesampleisrepresentativeandtheresultcanbegeneralized.SamplingprocessNonprobabilitysamplingTheinclusionorexclusionofelementsinasampleislefttothediscretionoftheresearcher.Noteveryelementofthetargetpopulationhasachanceofbeingselectedintothesample.Theprobabilitythataparticularunitwillbeincludedinthesampleisnotknown.SamplingprocessTypesofsamplingmethodsProbabilitysampling:SimplerandomSystematicStratifiedClusterMultistageNon-probabilitysampling:ConvenienceJudgementSnowballQuotaSamplingprocessDeterminesamplesizeHowmanyunitsshouldbeincludedinthesample?FormulasbasedonstatisticaltheorySamplesizebasedonruleofthumbPrevioussimilarstudiesResearcher’sownexperiencejudgmentSamplingprocessIndesigningasampleresearchershouldconsider:DefinitionofpopulationSizeofsampleRepresentationofthesampleGeneralize-abilityofthefindings.Calculatingasamplefrompilotstudyn=suggestedsamplesizeZ=levelofconfidenceexpressedinstandarderrorss=populationstandarddeviationE=acceptablelevelofsamplingerrorMcDaniel&Gates(2019)pp369-3731.96*1.961.39*1.390.1*0.1

Z2

s2

n=

E2

What’saconfidencelevel?Thelikelihoodinpercentagesthattheresultsarerepeatableandreal,notaflukeorrandomoccurence.Ifits95%confidencelevel,itmeansthere’sa95%chancesimilarresultswouldbeobtainedfromrepeattests.

What’sanacceptablelevelofsamplingerror?Theextenttowhichtheresearcheriswillingtotoleratethefindingsfromthesamplebeingdifferentfromtheentirepopulation(ifwedidruntheentirepopulation)BUT,keepinmindwearealwaystalkingprobabilitiesofeventshappening!ConfidenceLevelThelevelofconfidence(Z)andtheamountoferror(E)mustbesetbytheresearcher.Howconfidentdoyouneedtobethatthesamplerepresentsthecharacteristicsofthewholepopulation?(usuallyset95%)McDaniel&Gates(2019)pp369-373SourcesoferrorPotentialSourcesofErrorinResearchDesignsTotalerrorRandomsamplingerrorsNon-samplingerrorsResponseerrorsNon-responseerrorsResearchererrorsInterviewererrorsRespondenterrors(Malhotra&Birks,2019p74)LimitationsFigure14.2 Aclassificationofsamplingtechniques

Conditionsfavoringtheuseof:FactorsNon-probabilitysamplingProbabilitysamplingNatureofresearchExploratoryConclusiveMagnitudeoferrorsLargernon-samplingerrorsSamplingerrorsarelargerVariabilityinpopulationHomogenousHeterogenousStatisticalconsiderationsUnfavorableFavorableOperationallyFavorableunfavorableMalhotraBirks,3rdEu.Ed.,pg.421MeasurementsandScalingMeasurementandscaling:

Whatdotheymean?Measurement=theassignmentofnumbersorothersymbolstocharacteristicsofobjectsaccordingtocertainpre-specifiedrulese.g.assigning1,2or3inanattitudequestion1=Idon’tlikeit,2=notsureifIlikeitornot,3=IlikeitScaling=thegenerationofacontinuumuponwhichmeasuredobjectsarelocateddecidingthemagnitudeofdifferences132NotsureLikeitDon’tlikeit40SitNominalscaleAscalewhosenumbersserveonlyaslabelsortagsidentifying&classifyingobjectswithastrictone-to-onecorrespondencebetweenthenumbers&theobjectsNoorderofrankingE.g.doyoulikeshopping?

0=No,1=Yes41SitOrdinalscaleAscalethatallowstherespondentstorankastimulusobjectintermsofpreferences,importance,likingetcCandeterminewhetheranobjecthasmoreorlessofacharacteristicthanotherobjectBUT,notthemagnitudeofdifferenceRANKthesparetimeactivitiesintermsofyourpreferencebyassigningarankfrom1to5(1=mostpreferred,2=secondmostpreferredetc)shoppinghavingafewpintsattheChancellorplayingWiiorPlaystationupdatingmyFacebookorTwitter42SitIntervalscaleAscaleinwhichthenumbersareusedtorankobjectssuchthatnumericallyequaldistancesonthescaleequaldistancesinthecharacteristicbeingmeasuredShoppingismymostfavouritesparetimeactivity1=stronglydisagree2=disagree3=neitherdisagreeoragree4=agree5=strongagree43SitRatioscaleThehighestscale.Ascaletoidentifyorclassifyobjects,rankordertheobjects,andcompareintervalsordifferenceszeropointisfixedallowstocomputeratiosofscalevaluesSpecifytheamount(￡)spentatPrimarkinthepastthreemonths￡53.8044SitWhatiscomparativescaling?Involvesthedirectcomparisonofstimulusobjectse.g.doyoupreferiPhoneorBlackberry?Interpretedinrelativeterms&haveonlyordinalorrankorderpropertiesAlsoknownasnon-metricscaling45SitWhatisnon-comparativescaling?AlsoknownasmetricscalingEachobjectisscaledindependentlyoftheothersinthestimulussetTheresultingdatatendtobeintervalorratioscaledIndicatethedegreeofyourpreferenceforiPhoneona7-pointscaleNotatallpreferredGreatlypreferred46SitTypesofnon-comparativescalingSit47Non-comparativescalingContinuousratingscalesItemisedratingscalesLikertSemanticdifferentialStapelLikertscaleAratingscalethatrequirestherespondentstoindicateadegreeofagreementordisagreementwitheachofaseriesofstatementsaboutthestimulusobjectsMostcommonlyusedinmarketingresearchTypically5or7responseoptionsareusedrangingfrom‘stronglydisagree’to‘stronglyagree’48SitSemanticdifferentialscaleA7-pointscalewithendpointsassociatedwithbipolarlabelsthatsemanticmeaningScalescores-3to+3OR1to7BoringExcitingO2storeis...X49SitSemanticDifferentialExampleTheSemanticDifferentialScale

Thesemanticdifferentialscaleasksapersontorateaproduc